Location-Aware Pilot Allocation in Multicell Multiuser Massive MIMO Networks

Abstract

We propose a location-aware pilot allocation algorithm for a massive\nmultiple-input multiple-output (MIMO) network with high-mobility users, where\nthe wireless channels are subject to Rician fading. Pilot allocation in massive\nMIMO is a hard combinatorial problem and depends on the locations of users. As\nsuch, it is highly complex to achieve the optimal pilot allocation in real-time\nfor a network with high-mobility users. Against this background, we propose a\nlow-complexity pilot allocation algorithm, which exploits the behavior of\nline-of-sight (LOS) interference among the users and allocate the same pilot\nsequence to the users with small LOS interference. Our examination demonstrates\nthat our proposed algorithm significantly outperforms the existing algorithms,\neven with localization errors. Specifically, for the system considered in this\nwork, our proposed algorithm provides up to 37.26% improvement in sum spectral\nefficiency (SE) and improves the sum SE of the worst interference-affected\nusers by up to 2.57 bits/sec/Hz, as compared to the existing algorithms.\n